ElevenLabs Dubbing v2

Dubbing translates audio and video into more than 90 languages while keeping each speaker sounding like themselves. It preserves voice, emotion, timing, and tone, and keeps the original background audio, so you get a natural-sounding dub without re-recording anyone.

Give it a file or a URL, pick a target language, and you get back the dubbed audio track.

What it’s good for

• Reaching international audiences with content you already have
• Localizing videos, podcasts, and interviews without re-recording voice talent
• Keeping a speaker’s identity and delivery intact across languages

Inputs

• target_language — the BCP-47 language tag to dub into, for example es, fr, ja, or a region-qualified dialect like es-MX or pt-BR.
• source_url — a public URL to the source audio or video (a direct link, YouTube, TikTok, etc.).
• audio_or_video_file — upload a file instead of a URL. Provide either this or source_url, not both.
• source_language — the BCP-47 tag of the source media, for example en. Leave it as auto to detect the language automatically.
• cloning_strength — how strongly the dubbed speakers clone the source voices, from 0 to 10. Higher values prioritize resemblance to the original voice; lower values give the model more room for natural delivery in the target language. The default of 7 works well for most content.

Output

The dubbed audio track as a lossless FLAC file. Multiple speakers, overlapping speech, and background audio are handled automatically.

Languages

Dubbing supports 90+ languages, including English, Spanish, French, German, Japanese, Chinese, Arabic, Hindi, and many more. Some languages also support region-qualified dialects such as en-GB, es-MX, fr-CA, and pt-BR.

A note on quality

Dubbing v2 is in alpha, so you may hit the occasional rough edge. Free-tier dubs are watermarked; paid-tier dubs are not.